**3. Indoor airflow**

#### **3.1 Airflow temperature measurement test room**

An airflow temperature measurement test room was used to develop the indoor airflow control model. This test room is constructed with wooden walls, windows, and doors that mimic a room in a Japanese wooden house. The outside air temperature in this room can be controlled by an air conditioning system. This room is equipped with thermocouples, which can measure indoor airflow temperature distribution. **Figure 11** shows the layout of this room and the arrangement of the thermocouples. The air temperature was acquired using a test room with thermocouples taken down from the ceiling. As shown on the right of **Figure 11**, the temperature distribution of the airflow can be visualized by interpolating the temperature in the space without thermocouples through heat conduction calculations.

#### **3.2 Numerical model**

#### *3.2.1 Model development*

**Figure 12** shows how the mesh of the computational model is created. We simplified indoor unit cross-flow fans and vanes to an inlet boundary condition and indoor

*Application of CFD to Prediction of Heat Exchanger Temperature and Indoor Airflow Control… DOI: http://dx.doi.org/10.5772/intechopen.110076*

#### **Figure 11.**

*Schematic diagram of the thermocouple arrangement. The test room is 3.6 (m) deep by 7.2 (m) high by 2.4 (m) wide. Thermocouples are suspended from the ceiling of the test chamber, and thermocouples are placed at intervals of 0.05 to 0.4 (m). A total of 1134 thermocouples were installed in a layout of nine thermocouples in the x direction, 18 in the y direction, and seven pointing in the z direction (height).*

#### **Figure 12.**

*CFD mesh. This mesh consists of a room space, walls, windows, and a door. The room space is the indoor airflow environment, and wall indicates the floor, ceiling, and front, back, left side, and right side walls. The room air conditioner was installed on a back side wall. Outer surfaces of walls can set outside temperature as the boundary condition.*

unit inlet to outlet boundary condition. The boundary conditions for the inlet surface are (1) airflow volume rate, (2) airflow angle, and (3) airflow temperature. The walls, windows, and doors of the test room have small gaps through which drafts can enter and heat can leak out. It is difficult to model heat leakage from a draft numerically because air gap locations are unknown. Therefore, the initial temperature of the wall was set lower than the room temperature to model the heat leakage by the air gaps. The outer surface of the walls was set to the boundary condition of the outside air temperature in the experiment.

#### *3.2.2 Buyoyant modeling*

We use unsteady RANS model for indoor airflow control model. The governing equations were discretized by the finite volume method and solved using the SIMPLEC method. We used commercial CFD Code SCRYU/Tetra [12] for the calculation. To realize a rotating fan turbulence flow, the inlet condition was also given turbulence kinetic energy of 10 (m<sup>2</sup> /s2 ) and turbulence dissipation of turbulent kinetic energy of 0.1 (m<sup>2</sup> /s2 ). These values were obtained from the result of previous air-side

**Figure 13.**

*Temperature distribution. We measured 54 points of air temperature at 0.4 (m) in front of the air conditioner and compared two RANS models.*

CFD calculations. The blowing airflow from air outlet is acted as buoyant force, so that the Boussinesq approximation was used for buoyancy force as a momentum conservation equation.

To evaluate the influence of the turbulence model difference, we performed an experiment and two RANS model calculations. One model is linear RANS (standard kepsilon model) and the other is used nonlinear RANS [14]. **Figure 13** shows the results of an experiment and two calculations. The experimental result and calculation of the nonlinear turbulence model for temperature distribution show the long warm air region in a horizontal direction. Otherwise, linear turbulence model shows a vertical long hot air region. The standard k-epsilon model is isotropic turbulence and the Reynolds stresses act equally in all three directions in three-dimensional space. Since buoyancy forces act only in the vertical direction, they cannot be reproduced by an isotropic model. Therefore, we consider that the anisotropic model is closer to the experimental temperature distribution than the isotropic model. We selected a nonlinear turbulence model for investigating the influence of the buoyancy force.

**Table 3** is specification of CFD for our indoor air-control model.


**Table 3.** *Specification of CFD.*
